netbackup process monitoring

[richs24]

Hi,

i have been asked to provide a list of processes that we can monitor to determine whether or not our master/media servers are up and running.  Can anyone help me with this?

The issue i have, is that a number of the processes that run onthe master/media servers can spawn child processes tohelp with workload, and then close them down when it catches up.  I dont want to be reporting on these.

I am looking at getting NOM setup, but i have to set up this process monitoring in the meantime, so any help would be appreciated

regards

Rich

[rjrumfelt]

For the master:

http://seer.entsupport.symantec.com/docs/297548.htm

For a media server:

http://seer.entsupport.symantec.com/docs/297551.htm

[Andy_Welburn]

DOCUMENTATION: How to verify the required Veritas NetBackup (tm) 6.5 daemons / services are up and running on a master server
http://seer.entsupport.symantec.com/docs/297548.htm

DOCUMENTATION: How to verify the required Veritas NetBackup (tm) 6.5 daemons / services are up and running on a media server
http://seer.entsupport.symantec.com/docs/297551.htm

and for completeness:

DOCUMENTATION: How to verify the required Veritas NetBackup (tm) 6.5 daemons / services are up and running on a client
http://seer.entsupport.symantec.com/docs/297553.htm

***EDIT***
Touché  rj!

[richs24]

thanks, but a lot of these processes spawn child processes when the server is busy, are there any processes that i can set up monitoring against that dont spawn child processes?

[Andy_Welburn]

is busy" ?

A lot of these 'processes' will result in further processes when active & backups are running e.g. bptm, bpbrm etc. otherwise it wouldn't work!

Essentially the only relevance are the actual services & there are a limited number of these (per the original T/N's quoted).

May be worth looking through the following:
DOCUMENTATION: Details on the VERITAS NetBackup (tm) 6.0 Backup and Restore Process Flow. Covers all backup and restore processes from beginning to end.
http://seer.entsupport.symantec.com/docs/282015.htm

[jim_dalton]

Why not have your server up and running normally, without restore or backup running and look in the activity monitor->daemons and thats your list. Works for me every time.
Use something like serversalive and unix plugin to monitor if you dont like nom. I use both.

[richs24]

but my point is being missed

bprd and bpdbm are running multiple versions of themselves now on my system, and it is a quiet period now.  If i ever got to a stage where there were no backups (unlikely), i know there would only be 1 version of these 'services' running.
So my question is, are there any of these 'services' that dont spawn extras?

[Andy_Welburn]

any readily available info.

e.g the following is from the Admin Guide but it DOESN'T expand any further on the part(s) that I believe you are interested in:

"...
Stand-alone daemons: These NetBackup daemons always run and listen to accept connections. Examples include bpdbm, bprd, bpjobd, and vmd.
Multiprocess stand-alone daemons: NetBackup daemons that “fork” a child process to handle requests. Examples include bpdbm and bprd.
Single-process stand-alone daemons: NetBackup daemons that accept connections and handle requests in the same process.
inetd daemons: NetBackup daemons usually launched by inetd(1m)or bpinetd. Examples include bpcd, bpjava-msvc, and vnetd.
..."

This obviously appears to confirm what you don't want (i.e. bprd & bpdbm) but doesn't expand on what you do (single-process standalone daemons) - unless I really am missing the point......

[richs24]

no, thats my point.
at some stage i'll write a script to monitor which daemons appear and disappear, but until that point...........

[Anonymous]

(If UNIX servers) To go back to your original point what processes to monitor. I would have a look at the /etc/init.d/netbackup script - in there is all the scripts for the parent processes that are started and remain running for the duration and would be what I would be looking to check in your monitoring.

[Marianne · Accepted solution]

Main bprd and bpdbm have PPID of 1. Child/forked processes will have the main daemon PID as parent.

    root  5858     1   0   Jun 12 ?           1:59 /usr/openv/netbackup/bin/bpdbm
    root  5847     1   0   Jun 12 ?           3:27 /usr/openv/netbackup/bin/bprd

   root 25927  5847   0 14:35:09 ?           0:00 /usr/openv/netbackup/bin/bprd

 # ptree 5847
5847  /usr/openv/netbackup/bin/bprd
  25927 /usr/openv/netbackup/bin/bprd

[Will_Restore]

Netbackup processes rarely go down unless they are stopped/restarted manually (or the server is rebooted). 

[richs24]

and that is the answer i was looking for!
So obvious, i couldnt even see it!